After reading some articles [1] [2] yesterday about Docker and the "init"
process I got to thinking about the problems that we've been seeing on Travis.
The basic problem is that a Linux system may need an "init" process to work
properly when processes become zombies. Docker by default doesn't handle this
and the root process typically isn't an init process, so this can occasionally
cause quite a few problems.
We've been seeing spurious errors on Travis inside containers which look like
OOM and such, but my guess is that zombie processes were being reparented to the
top-level shell. The shell didn't expect the zombies and then behaved very
strangely.
This commit fixes these problems by using Yelp's "dumb-init" program [2] as the
init process in all of our containers. This ensures that there's a valid init
ready to reap children when they're reparented, which our test suite apparently
generates a bunch of throughout the tests and such.
[1]: https://blog.phusion.nl/2015/01/20/docker-and-the-pid-1-zombie-reaping-problem/
[2]: https://engineeringblog.yelp.com/2016/01/dumb-init-an-init-for-docker.html
This commit switches the default build system for Rust from the makefiles to
rustbuild. The rustbuild build system has been in development for almost a year
now and has become quite mature over time. This commit is an implementation of
the proposal on [internals] which slates deletion of the makefiles on
2016-01-02.
[internals]: https://internals.rust-lang.org/t/proposal-for-promoting-rustbuild-to-official-status/4368
This commit also updates various documentation in `README.md`,
`CONTRIBUTING.md`, `src/bootstrap/README.md`, and throughout the source code of
rustbuild itself.
Closes#37858
This commit configures our `.travis.yml` to test the full suite of tests we have
on Buildbot right now. A whole mess of docker images are added to the `src/ci`
directory which represent all the build environments for each configuration.
Each of these environments is then configured in `.travis.yml` to run on the
auto branch.
Note that the full matrix of tests aren't intended to be run on all PRs.
Instead, we continue to run only one entry in the matrix, forcing all others to
finish quickly. Only the `auto` branch should run the full matrix of builds.
Also note that the infrastructure hasn't quite been allocated yet to the
rust-lang/rust repository, so everything is disabled for now except for the one
build that happens on PRs. Once that infrastructure is allocated though we can
enable this and let it fly!
Notable modifications from the current test suite today:
* Android tests are run in rustbuild instead of the makefiles, for whatever
reason I couldn't get the makefiles to work on Travis.
* A debuginfo test was updated to work with the current version of the Android
NDK.
* Some dependencies in `mk/tests.mk` were fixed to allow running tests in
parallel.
